How to restore cache pool

[DBJordan]

I just set up an unraid 6.7.2 recently and assigned two nvme cache drives to the array to create a cache pool. One of the drives disappeared. Thinking it may be a heating issue, I powered off the server, waited for 30 minutes, and booted it again. It told me it could now see the drive and it wasn't part of the array, so I readded it to the cache pool and formatted it. Now my cache pool once again has both drives. (And I may have an overheating issue to address.)

 

However, my docker containers and VMs have disappeared. They were stored on the cache pool. I have two drives in the pool, so I should have some redundancy for them, but they're not showing up on the file system. I ran Mover, but that didn't restore them. How do I get their data restored from the pool's built in redundancy? I tried a few searches and didn't find anything, but maybe I'm just using the wrong terms. 

 

I'd post the logs, but really they're just full of docker and VM errors saying they can't find any of the directories or config data for the docker containers and VMs. Let me know if you think they'll be useful and I'll post them.

 

Any help is appreciated.

[trurl]

Go to Tools - Diagnostics and attach the complete diagnostics zip file to your NEXT post.

[JonathanM]

2 minutes ago, DBJordan said:

I readded it to the cache pool and formatted it.

The cache pool is one entity when it comes to format. You erased the data.

 

Possibly if you had asked for advice before adding the drive back things would be different, but since you formatted, I don't really see a way to recover.

 

For the future, there are plugins and scripts to facilitate regular backups of the cache pool to the array.

[DBJordan]

2 minutes ago, jonathanm said:

Possibly if you had asked for advice before adding the drive back things would be different, but since you formatted, I don't really see a way to recover.

Interesting. I can restore from backup. If the drive disappears again later, is there a way to recover the data since it's only 1/2 of the pool?

 

Diagnostics attached.

truesource-diagnostics-20191207-1454.zip

[itimpi]

It is worth pointing out that in 6.7.2 Unraid release there is a bug where the cache is not set up correctly to provide redundancy when using a cache pool.   This can be rectified by running the appropriate CLI command.    The bug is also rectified in the 6.8.0 RC series which look like becoming the new stable release very soon.

[JonathanM]

You need to fix the pool as described in this thread.

Also, don't immediately add the drive back if it drops, you have to remove both devices from the pool, start the array with no cache devices defined, stop the array, and add both back at the same time.

 

Redundancy isn't equivalent to backup, you still need file backups as you found out.

[DBJordan]

Thanks, everyone, for the replies and help.